Question

I am a Python beginner who is trying to get multiple random lines for a given category. The original file has three columns, but what I am interested is just one of those categories. The file(csv) looks like this:

   No,Size,Name
   10,1346,Cat
   24,423,Dog
   289,590,Cat
   12,302,Dog
   351,33,Cat
   51,812,Dog
   91,778,Cat
   1193,465,Cat
   44,178,Dog

None of the lines are identical and I want to get random 3 lines for each ‘Name’. This is what I have so far:

import random

with open('C:\Users\Owl\file.csv') as f:
    lines = f.readlines()[1:] #Skip heading

for line in lines:
    try:
        name = line[2]
    except:
        continue

for name in lines:
    for lines in random.sample(lines,3):
        print lines

f.close()

But I get something like this:

   12,302,Dog
   1193,465,Cat
   10,1346,Cat
   2
   3
   D

instead of something like this:

   1193,465,Cat
   10,1346,Cat
   91,778,Cat
   51,812,Dog
   44,178,Dog
   12,302,Dog

In the output I get now, I am not getting lines by ‘Name’ and somehow just letters/numbers after that. Then, I get “ValueError: sample larger than population” and terminates (actual file is much larger than the example here).

Also, if possible, is there an easy way to sort by “Name” in the output?

I have been struggling with this for hours looking it up on the Internet but have not been able to solve… Could anybody please help me? Thank you all!

Answer 1

You can do this much more easily by using itertools.groupby() and the csv module. We first make a csv.DictReader to give us easy access to the values, we then sort and group the list by the "Name" column, and then select the values.

import itertools
import csv
import operator
import random

with open("test.csv") as file:
    data = csv.DictReader(file)
    key = operator.itemgetter("Name")
    for name, items in itertools.groupby(sorted(data, key=key), key):
        print(name+":", random.sample(list(items), 3))

Which gives us:

Cat: [{'Size': '33', 'Name': 'Cat', 'No': '351'}, {'Size': '590', 'Name': 'Cat', 'No': '289'}, {'Size': '465', 'Name': 'Cat', 'No': '1193'}]
Dog: [{'Size': '178', 'Name': 'Dog', 'No': '44'}, {'Size': '812', 'Name': 'Dog', 'No': '51'}, {'Size': '302', 'Name': 'Dog', 'No': '12'}]

If you wish to make the dictionaries lists, it’s easy to do with a simple list comprehension:

[[item["No"], item["Size"], item["Name"]] for item in items]